PSI - Issue 77

P.D.A. da Silva et al. / Procedia Structural Integrity 77 (2026) 103–110 Silva et al. / Structural Integrity Procedia 00 (2026) 000–000

106

4

elements with triangular degradation law (COHAX4R). To replicate the applied E a , a mass was incorporated at one of the model’s edges. By applying a pre-estimated impact velocity to the mass, and considering its volume and density, it was possible to emulate the experimental E a as a kinetic energy instead. The adherends were modelled assuming elastic-plastic material behaviour, while the adhesive CZM elements employed a quadratic stress criterion to predict failure initiation. Damage propagation in the adhesive followed a linear energetic criterion. Boundary conditions were applied as shown in Fig. 2. Quadrilateral elements with a 0.2 mm size were employed, consistent with the chosen t a .

Fig. 2. Boundary conditions for the tubular joints.

3. Results 3.1. Validation with experiments

Validation was carried out by comparing the obtained results with experimental data and reference numerical values from reference (Valente et al. 2019). The adhesives AV138, DP8005, and XNR6852 E-2 were analysed by comparing P m derived from experiments and reference simulations with those obtained in this study (Fig. 3).

40

AV138

DP8005

XNR6852 E-2

30

26.92

28.39

24.00

14.43 19.93

19.90

20

P m [kN]

11.75 16.79

13.61

10

0

Experimental

Numerical reference

Current numerical

Fig. 3. Comparison between experimental and numerical P m values.

For the AV138, the numerical P m was 5.7% lower than the reference numerical values, while the deviation from the experimental data was 12.9%, which is within an acceptable range. In the case of DP8005, the numerical simulation provided results closely matching the reference P - δ curve, though some minor failure δ deviations were observed. The numerical P m showed a minimal difference of 0.2% from the numerical reference. However, the experimental P m was 15.6% lower than the numerical predictions, possibly due to challenges in physically replicating idealized simulation conditions. For the adhesive XNR6852 E-2, the numerical P m was 5.5% higher from the numerical reference and 16.3% higher than the experimental P m . Thus, the CZM was deemed sufficiently reliable for design.

3.2. Geometrical modifications

3.2.1. Outer chamfer The introduction of outer chamfers aims to correct the negative effect of the existing singularities due to the overlap end geometry. σ y and τ xy stresses are normalized over τ avg , which is the average τ xy stress for each joint geometry. Fig.